A Reliable, Self-Adaptive Face Identification Framework via Lyapunov Optimization

TL;DR

This paper introduces a queue-aware, self-adaptive face identification framework that dynamically adjusts sampling rates using Lyapunov optimization to balance performance and system reliability in resource-constrained environments.

Contribution

It presents a novel adaptive framework for real-time face identification that optimizes sampling rates to prevent queue overflow while maintaining high identification accuracy.

Findings

Effective in balancing identification accuracy and system reliability.

Preliminary simulation confirms the framework's effectiveness.

Adapts sampling rate dynamically based on queue status.

Abstract

Realtime face identification (FID) from a video feed is highly computation-intensive, and may exhaust computation resources if performed on a device with a limited amount of resources (e.g., a mobile device). In general, FID performs better when images are sampled at a higher rate, minimizing false negatives. However, performing it at an overwhelmingly high rate exposes the system to the risk of a queue overflow that hampers the system's reliability. This paper proposes a novel, queue-aware FID framework that adapts the sampling rate to maximize the FID performance while avoiding a queue overflow by implementing the Lyapunov optimization. A preliminary evaluation via a trace-based simulation confirms the effectiveness of the framework.